Combs and Sexual Selection in Black Grouse (Tetrao Tetrix)

Behavioral Ecology Vol. 11 No. 5: 465–471

Combs and sexual selection in black grouse (Tetrao tetrix)

Pekka T. Rintama¨ki,a Jacob Ho¨glund,a Eevi Karvonen,b Rauno V. Alatalo,c Niklas Bjo¨rklund,a Arne Lundberg,a Osmo Ra¨tti,d and Jyrki Voutia aDepartment of Population Biology and bDepartment of Animal Ecology, Evolutionary Biology Centre, Uppsala University, Norbyva¨gen 18 D SE-75236 Uppsala, Sweden, cUniversity of Jyva¨skyla¨, Department of Biological and Environmental Sciences, PO Box 35, FIN-40351, Jyva¨skyla¨, Finland, and dArctic Centre, University of Lapland, PO Box 122, FIN-96101 Rovaniemi, Finland

We studied supra-orbital combs in lekking black grouse (Tetrao tetrix) in relation to sexual selection at five leks in Finland 1991– 1998 and four leks in Sweden 1992–1995. Comb size was estimated in two ways: by observing its natural size in the field at different behaviors (‘‘observed comb size’’), and by measuring the comb size from captured birds (‘‘measured comb size’’). The size of combs is highly variable, and individuals can change it within seconds. Males express their larger combs during display, as compared to other behaviors. Observed mean comb sizes were larger on leks with a higher number of males and a higher number of copulations. Measured and observed comb sizes and copulatory success did not significantly correlate when all males where analyzed, but a positive and significant relationship between observed comb size and copulatory success was found within males that achieved copulations. Measured comb length correlated positively with the amount of testosterone. While females were present on the lek, displaying and successful males showed the largest observed comb size. When we compared observed comb size during fighting between successful and unsuccessful males and correlated comb size of pairs of fighting males with their fighting activity, no significant differences in comb size were found. The result that comb size correlated significantly with an increase in testosterone level and that larger comb size, within successful males, predicted higher copulatory success suggests that combs may be a cue for females to assess male quality. The lack of a significant relationship between observed comb size and fighting behavior suggests that comb size either has minor importance in male–male signaling on the lek or that males may express similar-sized combs during fighting to avoid serious fights and thus risk of comb injuries. Key words: black grouse, combs, copulatory success, sexual selection, Tetrao tetrix. [Behav Ecol 11:465–471 (2000)]

ntegumentary structures in the plumage or on parts of the good opportunities to observe and compare many male I body have evoked the interest in their function in sexual combs simultaneously. signaling since Darwin. Such ornaments may have evolved Comb and wattle ornaments are usually more conspicuous through intra- and/or intersexual selection (i.e., they are sig- in males than in females (Owens and Short, 1995), which nals between males or help females to assess males; Zuk, may indicate their function as ornaments. Expression of 1991). Supra-orbital combs and wattles are fairly common in combs and wattles is directly connected to androgen pro- gallinaceous birds, but they also occur in other taxonomically duction (i.e., testosterone; Ligon et al., 1990; Moss et al., unrelated bird families (Holder and Montgomerie, 1993a). 1979; Owens and Short, 1995; Zuk et al., 1995), whereas Brightly colored skin parts are particularly common in the feather ornament size seldom depends on current levels of Tetraonids ( Johnsgard, 1983), and their function in a sexual testosterone secretion (Stokkan et al., 1988; Witschi, 1961). selection context has mainly been studied in nonlekking It is also possible that androgens not only increase the ex- grouse species, both in the context of male–male signaling pression of sexual ornaments, but also suppress the immune (Allee et al., 1939; Graves et al., 1985; Ligon et al., 1990; Moss function (Folstad and Karter, 1992). Thus, combs may be et al., 1979; Myhre, 1980) and female choice (Brodsky, 1988; particularly useful as honest signals for prospecting females, Hannon and Eason, 1995; Holder and Montgomerie, 1993a; if these traits provide females with accurate information on Zuk et al., 1990b). However, few studies of the importance of current male condition. For example, in red jungle fowl fleshy structures have been done in lekking species (Gratson (Gallus gallus), the size of combs and wattles may reflect et al., 1991; see also Johnson and Boyce, 1991; Spurrier et al., individual variation in parasitemia, disease resistance, nutri- 1991). This is somewhat surprising because many lekking tional status, and general health (Allee et al., 1939; Johnson grouse show conspicuous structures of this kind during dis- et al., 1993; Zuk et al., 1990a). An alternative, or additional play. Furthermore, the possibility of combs serving as signals explanation for the occurrence of combs and wattles is that is especially plausible in lekking species because males on the they may be important in male–male aggressive signaling lek form tightly clumped aggregations where combs may be (Zuk, 1991). Studies in red grouse (Lagopus lagopus scoti- observed and compared at a close distance by conspecifics. cus), red jungle fowl, and rock ptarmigan (Lagopus mutus) Thus, both choosy females and competing males should have have indeed suggested that variation in the size of fleshy structures are linked to variations in levels of aggression (Al- lee et al., 1939; Graves et al., 1985; Holder and Montgomer- Address correspondence to P. T.Rintama¨ki. E-mail: pekka. ie, 1993a; Moss et al., 1979). [email protected]. In the present study we examined whether comb size was Received 2 June 1999; revised 15 October 1999; accepted 17 De- related to testosterone levels, to a suite of behavioral and ter- cember 1999. ritorial traits, and to copulation success. We also investigated 2000 International Society for Behavioral Ecology whether comb size differed among rivaling males. 466 Behavioral Ecology Vol. 11 No. 5

Figure 1 Comb drawings from the front of a bird (rookooing position) showing the rank at which comb size was estimated. Observation procedure of combs on a five-point rank: 1, not swollen, the combs appear only as a thin red line above the eye; 2, minor swelling, there is a considerable gap between the combs, 3, swollen but the space on the top of the head can be seen clearly; 4, heavily swollen but the feathers on top of the head can just be seen, 5, wholly erected, both combs are in contact, and feathers on the top of the head between the combs are not visible. Scale, 1:2.

METHODS only resident males that were regularly attending the lek (25% attendance or more in relation to the most attendant male) Data collection because males with lower values were often nonterritorial ju- During 1992–1998 we gathered data on observed comb size veniles or males that had unstable territories. These males from four black grouse leks in central Sweden and from one were difficult to identify during the 1- to 2-min scan sampling in central Finland. In Sweden one lek was studied in 1992, of the resident males. Eight males were occasionally seen and 1993, and 1995 (lek no. 1), another in 1993 and 1995 (lek omitted (11% of the males included in the study), with a no. 2), and two in 1997 (leks no. 4 and 5); the data from mean Ϯ SD attendance rate of 11% Ϯ 7%. Finland derives from one lek in 1995, 1997, and 1998 (lek no. 3). For details of the study areas and general methods, see Combs Rintama¨ki et al. (1995a,b, 1999) and Ho¨glund et al. (1997). All leks used in observed comb size analyses were situated on The red combs of grouse are distinct during the mating sea- open bogs, and lek size varied from 3 to 13 resident males son (Gjesdal, 1977), and males are able to change comb size (mean Ϯ SD, 7 Ϯ 3 males). In Finland, the majority of the rapidly (in a few seconds; Rintma¨ki et al., personal observa- males (77%) were color banded, enabling us to recognize the tions from captured males) by controlling blood circulation birds individually; in Sweden only a few males were ringed. in the comb tissue (Holder and Montgomerie, 1993a). How- Resident males (see below) on leks are very faithful to specific ever, although males can control comb size, they cannot hide territories and show individual plumage variation, allowing their combs completely. Because the size of the comb of cap- identification of individuals. To avoid replication of data on tured males does not provide accurate information about how individuals, we included only the first year of observation for males express combs on the lek, we developed a rank system males present at a lek in more than 1 year. Leks were observed (1 to 5, observational comb size; see Figure 1 for explana- with binoculars (8–10ϫ) and telescopes (20–60ϫ) from hides tions) by which we estimated each male’s comb size when positioned close to leks (10–40 m to males). scan-sampling the males. In addition to observed comb size estimates, we measured combs of 17 adult males captured dur- Male characteristics ing the mating season at five leks in 1991 in Finland. Different studies of grouse have used different methods to estimate For each male, we used the number of observed copulations comb size (see, e.g., Brodsky, 1988; Holder and Montgomerie, as an estimate of his total copulatory success. In black grouse, 1993a; Zuk et al., 1992). As in the red jungle fowl studies (e.g., females usually copulate only once, if a copulation is not dis- Chappell et al., 1997) we believed comb length to be the most rupted, and DNA fingerprinting has confirmed that the male reliable measure because the length appears to change less seen copulating sires all chicks in the clutch (Alatalo et al., rapidly as compared to vertical comb height. We measured 1996a). We examined male behaviors, territory characteristics, the maximum nonflattened horizontal comb length (to the and the number of males at the lek in the context of both nearest 0.1 mm) from one side of the head by using digital measured and observed comb size (these measures are de- calipers. fined below). Behavior was divided into three main categories: Observational comb size estimates were taken when females display (rookooing; the principal vocal display, flutter jump, were both present and absent from the lek (leks 3–5 only). and hissing), fighting, and other activities (standing still, walk- On leks 1–2, the low number of female visits prevented us ing, preening, eating). We also noted individual differences from using observations from these leks in the analyses. Black in rates of lek attendance. Activities were calculated as pro- grouse males respond to the presence of females by increasing portions of how much time each male spent in each activity their activity (Ho¨glund et al., 1997), and we therefore checked in relation to other activities. Rate of attendance was calculat- whether comb size was affected by female attendance. ed as presence on the lek in relation to the male(s) with the highest attendance; the male with the highest attendance was Reliability of observational comb estimates given a score of 100%. During scan sampling, we plotted each male’s location to the closest 1 m on the maps using a 10 ϫ In spite of detailed instructions to observers, the observed 10-m grid system on each lek. Territory location was then de- comb size is likely to be more prone to errors among observ- fined in terms of the distance from the territory center to the ers than are the direct measurements. To see how reliable our lek center (median of all male positions), and territory size observed comb estimates were, we compared between-observ- was calculated as the diameter (meters) of an ellipse within er variation at lek 3. At a given moment two observers inde- which 90% of the observation points were situated (average pendently ranked a specific male comb size. The observation for horizontal and vertical diameters). This method excludes procedure was repeated 25 times on a total of 5 mornings occasional positions outside a male’s territory. We included including 10 different males and 7 behaviors. Of 25 comb Rintama¨ki et al. • Combs and sexual selection 467

Table 1 Table 2 Mean observed comb size of all males during female absence and Correlations between male observed comb size and behaviors presence (fighting, rookooing) and lek attributes (distance to the lek center, territory size, and attendance) Females Females Trait absent n present n Females Females not present present All behaviors 3.56 (0.02) 1456 4.03 (0.04) 494 Trait rpnrpn Display 3.92 (0.04) 639 4.18 (0.05) 268 Fighting 3.77 (0.04) 392 3.90 (0.05) 182 Other 3.05 (0.04) 308 3.35 (0.08) 31 Observed comb size Fighting .08 .34 73 Ϫ.14 .23 37 Rookooing .12 .13 73 .12 .37 37 The data contain only leks where we had collected data on males Attendance .08 .35 71 Ϫ.18 .14 35 both during female presence and absence. Mean observed comb Distance to lek center .17 .04 73 Ϫ.07 .55 37 size (Ϯ SE) during all behaviors, displaying, fighting, and other Territory size Ϫ.19 .02 71 Ϫ.07 .54 37 behaviors are indicated. Measured comb size Fighting .02 .95 16 estimates, 23 were identical, and both the observers obtained Rookooing .03 .92 16 a mean rank of 4.0 (SDs for observers were 0.58 and 0.65) Attendance Ϫ.05 .83 16 Ϫ with identical minimum and maximum observed comb size Distance to lek center .17 .52 16 ranks of 3.0 and 5.0. The result was also significantly repeat- Territory size .35 .22 14 able (Lessels and Boag, 1987; R ϭ .89, F ϭ 17.7, p Ͻ .001). In the two cases when observers’ estimates differed, the esti- Data of observational comb sizes were analyzed using Kendall rank mate difference was one rank. Therefore we consider that es- correlation analysis, and data of measured comb sizes were analyzed with randomization (resampling) test (5000 simulations). Observed timated observational comb size is accurate enough to detect comb size analyses are divided into two groups (i.e., observations differences between males and behaviors (note also sample when females were present or not). All the results were sizes and SEs given in Table 1). To correct for lek site and nonsignificant after correction for multiple tests. study year and to minimize possible differences among ob- servers, we used Kendall’s nonparametric partial rank order correlation analysis. dioimmunoassay after extraction and partial purification on diatomaceous earth-glycol microcolumns. To study the rela- Comb size during different behaviors and territorial tionship between measured comb size and testosterone levels, features, the relationship with copulatory success, and the we correlated comb horizontal length with testosterone levels. difference between fighting pairs of males We used one-tailed p values because the expectation based on earlier studies (see above) was that larger comb size is related To study how observed comb size was related to different be- to higher testosterone titers. haviors and territorial features, we first performed correla- tions of observed and measured comb sizes with different male characteristics (Table 2). Thereafter we correlated mea- Statistical analyses sured and observed comb sizes during different behaviors We used both parametric and nonparametric tests to analyze with the total copulatory success of each male (Table 3). Fi- our data. All tests are two-tailed unless otherwise stated. Re- nally, to study observed combs and the male–male signaling sampling (randomization) tests were done by simulating orig- hypothesis, we performed two tests. We first looked at differ- inal values 5000 times. When more than one variable was in- ences in observed comb size between different fighting pairs cluded in the analyses, we corrected the p values for multiple of males that had different copulatory success using paired t tests (Ury, 1976). Means include SD or SE values. tests. If successful males show larger observed combs during fighting, this might imply that combs are used by males to signal dominance relationships (Graves et al., 1985; Holder RESULTS and Montgomerie, 1993a; Moss et al., 1979). Second, if there Comb expression is no significant difference between observed comb sizes of fighting males with different copulatory success, combs may Males showed the largest combs when rookooing (mean Ϯ not be important in signaling dominance between males, or SD ϭ 3.7 Ϯ 0.6, n ϭ 73 males) and fighting (3.5 Ϯ 0.5, n ϭ alternatively, signaling similar-sized combs may prevent fights 65), with a significant difference between these behaviors that could cause comb injuries. To see if the comb sizes of (Wilcoxon matched-pairs signed-rank test, t ϭ 4.05, n ϭ 61, pairs of fighting males depended on how often they fought p Ͻ .001). When males were inactive (engaged in ‘‘other ac- with each other, we correlated comb size differences of these tivities’’), their comb sizes were clearly smaller (2.8 Ϯ 0.8, n male pairs during fighting with the their fighting activity (the ϭ 69) than when fighting (t ϭ 6.25, n ϭ 59, p Ͻ .001) and number of fights by each male pair divided by the total num- rookooing (t ϭ 7.02, n ϭ 69, p Ͻ .001). Comb size estimates ber of fights of all male pairs). showed the same direction during female presence, although observed comb sizes were higher for all behaviors when fe- males were present (Table 1). Measured comb and testosterone The mean observed comb size was larger on larger leks The testosterone measurement procedure has been reported than on smaller leks (females not present; Spearman rank ϭ ϭ Ͻ elsewhere (Alatalo et al., 1996b). Briefly, we captured 17 males order correlation: rs .68, n 10 leks, p .05). The rela- during the mating season between 16 April and 7 May. Blood tionship was even more clear when we correlated observed ϭ ϭ was collected from the basilic wing vein, and plasma was sep- mean comb size during display with lek size (rs .71, n 10, Ͻ ϭ ϭ Ͻ arated with a centrifuge (10 min at 10,000 rpm). Plasma levels p .01, Figure 2) and fighting (rs .77, n 10, p .01), ϭ ϭ of testosterone were measured (in duplicate) by a single ra- but not in the case of other activities (rs .18, n 10, ns). 468 Behavioral Ecology Vol. 11 No. 5

Table 3 Males’ observed comb sizes and measured comb sizes in relation to copulatory success

Females not present Females present rnp rnp

Observed comb size Total all males .01 73 .91 Ϫ.07 37 .56 Total successful males .32 29 .02 .43 20 .01* Rookoo, all males .03 73 .75 .13 30 .35 Rookoo, successful males .30 29 .04 .63 16 .001** Fight, all males Ϫ.10 66 .30 .01 21 .95 Fight, successful males .25 27 .09 .41 12 .07 Other, all males .19 69 .04 .32 16 .14 Other, successful males .38 28 .08 .44 7 .19 Measured comb size All males .23 17 .41 Successful males .56 9 .16

Successful (more than one copulation) and unsuccessful males (no copulations) were analyzed separately for females present or absent on the lek. Total observed comb sizes (all observations included) while rookooing (rookoo), fighting (fight), and engaged in other activities (other) are means of a male’s observed comb rank during these different behaviors (Kendall rank order correlations). Measured comb size in relation to copulatory success was analyzed separately among successful and unsuccessful males, and analyses were done using randomization (resampling) tests. Measured comb size is the horizontal comb length. Analyses account for lek site and study year. Two- tailed significances after correction for multiple tests are *p Ͻ .05 and **p Ͻ .01.

Positive and significant correlations between observed comb Combs, male characteristics, and testosterone size and lek size may suggest that males with larger combs Table 2 summarizes correlations between male mean ob- prefer larger leks. That observed comb size during other ac- served and measured comb size and different behaviors and tivities was much more weakly correlated with lek size than lek attributes. After correcting for multiple tests, none of the other behaviors suggests that this is not the case. However, to characteristics correlated significantly with male observed test this directly, we compared male observed comb sizes from comb size, irrespective of female presence. Further, measured which we had information from 2 years at different lek sizes. comb size was not significantly correlated with male territorial We then found that eight out of nine males had, on average, features and behavioral characteristics. Measured comb 0.5 smaller (mean Ϯ SD ϭϪ0.45 Ϯ 0.39) observed comb length was significantly correlated with log-transformed tes- ranking when displaying on smaller leks than on larger leks tosterone levels, both in all males (r ϭ .45, n ϭ 17, p ϭ .04), (sign test, p Ͻ .02). and within successful males (r ϭ .58, n ϭ 9, p ϭ .05).

Comb size and female presence We examined the effect of female presence on comb size on one lek (lek 3, years 1995, 1997, and 1998) by following the comb sizes of the same males (n ϭ 22 different males includ- ing 5 nonterritorial individuals) when females were not pre- sent on the lek and when one or more females were present. The interval between subsequent observations was Յ 10 min, allowing us to detect male responses in comb size in relation to female presence. The observed mean comb size (mean Ϯ SD) during all behaviors (display, fighting, other) was signif- icantly affected by the presence of females (comb size, females not present: 3.84 Ϯ 0.68; females present: 4.02 Ϯ 0.70; Wil- coxon signed-rank test: t ϭ 4.21, n ϭ 163, p Ͻ .001, where mean ϭ comb size according to a rank 1–5). The result was the same despite male behavior (display, females not present: 4.08 Ϯ 0.71 and present: 4.30 Ϯ 0.68, t ϭ 3.05, n ϭ 79, p ϭ .002, fighting, females not present: 3.63 Ϯ 0.57 and present: 3.78 Ϯ 0.62, t ϭ 2.6, n ϭ 73, p ϭ .009). Because female pres- ence influences male comb size, we analyzed our data both when females were present and when they were not. Figure 2 Mean observed comb size of black grouse during display (rookooing) in relation to lek size. The data were collected when Combs and copulatory success no females were present on the lek. Different leks are indicated with different symbols: triangles, lek 1; diamonds, lek 2; circles, lek Mean observed comb size of all males on leks correlated sig- 3; square, lek 4; horizontal bar, lek 5. nificantly with the number of copulations on the lek (Spear- Rintama¨ki et al. • Combs and sexual selection 469

Figure 4 Mean male comb size among successful males (more than one Figure 3 copulation) during display (rookooing) in relation to male Mean observed comb size in relation to the number of copulations copulatory success. Data were collected when females were not on leks. Different leks are indicated with different symbols: present on the lek. triangles, lek 1; diamonds, lek 2; circles, lek 3; square, lek 4; horizontal bar, lek 5. per fighting male pair). When we correlated the observed comb size difference of each male pair with their fighting man rank order correlation: r ϭ .88, n ϭ 10 leks, p Ͻ .01, s activity, we did not find a significant relationship (all males Figure 3). The relationship was positive, but nonsignificant, ϭ ϭ ϭ included: rs .07, n 89, ns, and only successful males: rs when we correlated mean measured comb size and the num- ϭ ϭ ϭ .05, n 22, ns). ber of copulations (rs .32, n 4 leks, ns), probably because we did not have information on measured comb sizes of all males and because of low statistical power due to few leks. DISCUSSION When all males where included in the analysis, mean observed Combs and female choice comb size, comb size during rookooing, fighting, and other activities were not significantly correlated with male copula- It has been suggested that individuals have to trade resources tory success, either when females were present or absent (Ta- allocated to maintain immunological defense against an in- ble 3). During female presence at a lek with only successful creased expression of sexual ornaments (Folstad and Karter, males (Ն 1 copulation) included in the analyses, however, 1992; Sheldon and Verhulst, 1996). The hypothesis predicts males with the largest mean combs and largest combs during that hormones cause the expression of sexual ornaments rookooing enjoyed the highest copulatory success (Table 3, while suppressing immune function, and that the costs of an Figure 4), although the relationship was not clearly linear. ornament would be lower for individuals in better physiolog- There seems to be a threshold within successful males; those ical condition (Nur and Hasson, 1984; Sheldon and Verhulst, with observed comb size Ն4 during display seem to enjoy 1996). In willow ptarmigan and red jungle fowl, comb size higher copulatory success (Figure 4). Measured comb size was depends on individual differences in testosterone levels (Stok- not significantly correlated with copulatory success within suc- kan, 1979; Zuk et al., 1995), and red jungle fowl males (e.g., cessful males, although the association was in the same direc- Ligon and Zwartjes, 1995a,b; Ligon et al., 1998; Zuk et al., tion as that obtained from observed comb size analysis (Table 1990b) as well as ptarmigan males (Brodsky, 1988; Hannon 3). and Dobush, 1997; Hannon and Eason, 1995) with larger combs tend to enjoy higher copulation success than males with smaller combs. High testosterone levels clearly are asso- Comb size and male–male signaling ciated with black grouse copulatory success (Alatalo et al., The mean difference in comb rank between two fighting 1996b), and we here add the finding that measured comb males was on average 0.12 when we included only successful length correlates positively with male testosterone levels. In males in the analyses (in the paired test the observed comb addition, larger observed comb size predicts higher copula- size during fighting of more successful males was compared tory success among successful black grouse males. Hence, re- to that of less successful ones, t ϭ 0.91, df ϭ 21, p ϭ .37, n sults obtained from both nonlekking and lekking species gen- ϭ 3.3 Ϯ 3.3 comb observations per fighting male pair). In the erally fit with the expectations of the immunocompetence previous analysis more successful males had, on average, handicap hypothesis and suggest that females may use comb slightly larger combs. When we added into the previous anal- size to evaluate male current condition. ysis male pairs that had higher copulatory success versus no In black grouse, using the data of all males, however, cop- success, the mean difference in comb rank during fighting ulatory success was not associated with comb size, either when was on average Ϫ0.07 (t ϭϪ0.66, df ϭ 63, p ϭ .51, n ϭ 4.7 measured at capture or when observed in undisturbed birds Ϯ 6.0 comb observations per fighting male pair). Finally, un- participating at the lek. A similar lack of significant association successful (copulation ϭ 0) males showed the least difference between comb size and copulatory success has also been in comb size during fighting (mean difference ϭϪ0.05, t ϭ found in another lekking grouse species (e.g., the sharp-tailed Ϫ0.29, df ϭ 24, p ϭ .77, n ϭ 3.4 Ϯ 5.2 comb observations grouse, Tympanuchus phasianellus; Gratson, 1993; Gratson et 470 Behavioral Ecology Vol. 11 No. 5 al., 1991). The distribution of copulations between males is cessful than the other in achieving copulations and also in typically skewed on leks, with only a few of the males achieving situations between pairs of males that had no success. The many copulations (Alatalo et al., 1992; Ho¨glund and Alatalo, results suggested that observed comb size did not differ sig- 1995; Kokko and Lindstro¨m, 1997; Widemo and Owens, nificantly between fighting males, either when comparing 1995). In the leks included in this study, 64% (SD ϭ 18.4) of males with different success rates or when comparing males the males had no success at all, while the most successful male that had no success. In addition, male pairs that fought more on a single lek achieved, on average, 63% (SD ϭ 22.5) of the in relation to other male pairs did not seem to differ in comb copulations. In the black grouse, as well as in lekking species size. in general, successful males settle in central territories (Fiske During fights, males seem to aim to damage the comb of et al., 1998). Thus it is possible that particular males, depend- the opponent (Owens and Short, 1995), and it has been sug- ing on other characteristics such as territory history (Kokko gested that because combs are frequently attacked during et al., 1998, 1999) and age (Kokko, 1997), may be rejected as male–male interactions, the condition of these ornaments mates whatever comb size they are able to show. As an ex- may honestly advertise male fighting ability (Holder and ample, a new adult male displaying in the periphery at lek 3 Montgomerie, 1993a). Because females show a preference for showed a nearly maximum mean observed comb size of 4.95 combs and males try to hit other males’ combs, we suggest but failed to mate with any female, whereas the two top males that showing less than maximally erected combs during fights near lek center both achieved 19 copulations and had mean may be a way to avoid injuries and thus keep the combs in observed comb ranks of 3.53 and 3.49. Given that females are good condition for females to evaluate. unlikely to base their choice solely on comb size, we used only the successful males in the analysis (more than one copula- tion) and found that males with larger observed comb sizes Conclusions: sexual selection were more successful in achieving copulations. The result was Exaggerated secondary sexual characters, such as combs and in the same direction but not significant among measured wattles, may provide sampling females reliable cues to use as males. Moreover, observed comb size during female presence honest signals of male quality. We did find support for this showed a stronger relationship with male copulatory success idea; black grouse females preferred males with larger ob- than did observed comb sizes when the females were not eval- served combs (and a similar trend for measured combs), al- uating males. though the result was significant when the analysis was based The relationship between comb size and copulatory success on successful males only. Our results that black grouse males suggests that male comb size may be an honest signal of male showed similar size combs during fights regardless of their quality for black grouse females. How, then, can we explain copulatory success and fighting intensity may suggest that the positive relationships between observed comb size and lek males ‘‘hide’’ their combs during fights in order to protect size as well as the relationship between comb size and the them (Holder and Montgomerie, 1993a). In conclusion, we number of copulations? High quality males are more likely to have found evidence that females may use comb size to eval- join larger leks, which black grouse females also seem to pre- uate males, but also that males diminish their comb sizes in fer (Alatalo et al., 1992). However, our result that the same male–male interactions, probably to keep them in good con- males seem to show smaller combs when participating in a dition. Hence we suggest that the most likely explanation for smaller lek does not support this idea. Also, male comb size the existence of combs is female preference for them. was affected by female presence. We therefore suggest that both the number of males and the presence of females induce male testosterone secretion, which, in turn, affects comb size The data were collected with the help of Matti Halonen, Matti Hovi, Esa Huhta, Jukka Jokima¨ki, and Pirkko Siikama¨ki. We also thank the (but see also Alatalo et al., 1996b; Wingfield et al., 1990). Finnish Ringing Centre for its cooperation. We received helpful com- Interestingly, the number of copulations was much lower on ments on earlier versions of the manuscript from Peder Fiske, Staffan Swedish leks. This may be partly due to the lower population Ulfstrand, Marlene Zuk, and two anonymous referees. R.V.A and O.R. density in the Swedish study area (Ho¨glund and Sto¨hr, 1996; were supported by the Finnish Academy, P.T.R. by the Emil Aaltonen Lindstro¨m, 1994). foundation and Nordisk forskerutdanningsakademi (NorFA), E.K. by NorFA, and J.H. and A.L. by the Swedish Natural Sciences Research Council. Combs and male–male signaling

Another reason apart from female choice for the evolution of REFERENCES combs is that they may be important signals in interactions between males (Hansen and Rohwer, 1986; Holder and Mont- Alatalo RV, Burke T, Dann J, Hanotte O, Ho¨glund J, Lundberg A, gomerie, 1993a,b; Zuk, 1991). This explanation could be par- Moss R, Rintama¨ki PT, 1996a. Paternity, copulation disturbance and ticularly pertinent on leks, where males are tightly clustered female choice in lekking black grouse. Anim Behav 52:861–873. and territories are adjacent to each other. Hence, males may Alatalo RV, Ho¨glund J, Lundberg A, Rintama¨ki PT, Silverin B, 1996b. Testosterone and male mating success on the black grouse leks. use combs to signal their intentions to other males. Observed Proc R Soc Lond B 263:1697–1702. comb size clearly differed depending on behavior, being Alatalo RV, Ho¨glund J, Lundberg A, Sutherland WJ, 1992. Evolution smaller when males were inactive, medium sized when they of black grouse leks: female preferences benefit males in larger were fighting, and largest during display. This suggests that leks. Behav Ecol 3:53–59. males indeed control their comb size according to their be- Allee WC, Collias NE, Lutherman CZ, 1939. Modification of the social havior. order in flocks of hens by the injection of testosterone propionate. Similar to our findings, no significant effects of larger comb Physiol Zool 12:412–440. size on increased fighting behavior were found in the sharp- Brodsky LM, 1988. Ornament size influences mating success in male tailed grouse (Gratson, 1993), which is also a lekking species. rock ptarmigan. Anim Behav 36:662–667. Chappell MA, Zuk M, Johnsen TS, Kwan TH, 1997. Mate choice and To test directly whether black grouse males use their combs aerobic capacity in red jungle fowl. Behaviour 134:511–529. during fighting, to show dominance, or, on the other hand, Fiske P, Rintama¨ki PT, Karvonen E, 1998. Mating success in lekking whether they hide their combs (at such situations combs may males: a meta-analysis. Behav Ecol 9:328–338. be of similar sizes), we compared comb sizes during fighting Folstad I, Karter AJ, 1992. Parasites, bright males, and the immuno- between pairs of two opponents, one of which was more suc- competence handicap. Am Nat 139:603–622. Rintama¨ki et al. • Combs and sexual selection 471

Gjesdal A˚ , 1977. External markers of social rank in willow ptarmigan. ritoriality and female choice on black grouse leks. Anim Behav 49: Condor 79:279–281. 759–767. Gratson MW, 1993. Sexual selection for increased male courtship and Rintama¨ki PT, Alatalo RV, Ho¨glund J, Lundberg A, 1995b. Mate sam- acoustic signals and against large male size at sharp-tailed grouse pling behaviour of black grouse females (Tetrao tetrix). Behav Ecol leks. Evolution 47:691–696. Sociobiol 37:209–215. Gratson MW, Gratson GK, Bergerud AT, 1991. Male dominance and Rintama¨ki PT, Karvonen E, Alatalo RV, Lundberg A 1999. Why do copulation disruption do not explain variance in male mating suc- black grouse males gather on leks outside the breeding season? J cess on sharp-tailed grouse (Tympanuchus phasianellus). Behaviour Avian Biol 30:359–366. 118:187–213. Sheldon BC, Verhulst S, 1996. Ecological immunology: costly parasite Graves HB, Hable CP, Jenkins TH, 1985. Sexual selection in Gallus: defences and trade-offs in evolutionary ecology. Trends Ecol Evol effects of morphology and dominance on female spatial behavior. 11:317–321. Behav Proc 11:189–197. Spurrier MF, Boyce MS, Manly BFJ, 1991. Effects of parasites on mate Hannon SJ, Dobush G, 1997. Pairing status of male willow ptarmigan: choice by captive sage grouse. In: Bird-parasite interactions (Loye is polygyny costly to males? Anim Behav 53:369–380. JE, Zuk M, eds). New York: Oxford University Press; 389–398. Hannon SJ, Eason P, 1995. Colour bands, combs and coverable badges Stokkan KA, 1979. Testosterone and daylength-dependent develop- in willow ptarmigan. Anim Behav 49:53–62. ment of comb size and breeding plumage of male willow ptarmigan Hansen SJ, Rohwer S, 1986. Coverable badges and resource defence (Lagopus lagopus lagopus). Auk 96:106–115. in birds. Anim Behav 34:69–76. Stokkan KA, Sharps PJ, Dunn IC, Lea RW, 1988. Endocrine changes Ho¨glund J, Alatalo RV, 1995. Leks. Princeton, New Jersey: Princeton in photostimulated willow ptarmigan (Lagopus lagopus lagopus) and University Press. svalbard ptarmigan (Lagopus mutus hyperboreus). Gen Comp En- Ho¨glund J, Johansson T, Pelabon C, 1997. Behaviourally mediated docrinol 70:169–177. sexual selection differences between successful male black grouse. Ury HK, 1976. A comparison of four procedures for multiple com- Anim Behav 54:255–264. parisons among means (pairwise contrasts) for arbitrary sample siz- es. Technometrics 18:89–97. Ho¨glund J, Sto¨hr S, 1996. A non-lekking population of black grouse Widemo F, Owens IPF, 1995. Lek size, male mating skew and the Tetrao tetrix. J Avian Biol 28:184–187. evolution of lekking. Nature 373:148–151. Holder K, Montgomerie R, 1993a. Context and consequences of Wingfield JC, Hegner RE, Dufty AM Jr, Ball GF, 1990. The ’challenge comb displays by male rock ptarmigan. Anim Behav 45:457–470. hypothesis,’ theoretical implications for patterns of testosterone se- Holder K, Montgomerie R, 1993b. Red colour bands do not improve cretion, mating systems, and breeding strategies. Am Nat 136:829– the mating success of male rock ptarmigan. Ornis Scand 24:53–58. 846. Johnsgard PA, 1983. The grouse of the world. Croom Helm, London. Witschi E, 1961. Sex and secondary sexual characters. In: Biology and Johnson LL, Boyce MS, 1991. Female choice of males with low para- comparative physiology of birds, vol. 2 (Marshall AJ, ed). New York: site loads in sage grouse. In: Bird-parasite interactions (Loye JE, Academic Press; 115–168. Zuk M, eds). New York: Oxford University Press; 377–388. Zuk M, 1991. Sexual ornaments as animal signals. Trends Ecol Evol Johnson K, Thornhill R, Ligon JD, Zuk M, 1993. The direction of 6:228–231. mothers’ and daughters’ preferences and the heritability of male Zuk M, Johnsen TS, Maclarty T, 1995. Endocrine-immune interac- ornaments in red jungle fowl (Gallus gallus). Behav Ecol 4:254– tions, ornaments and mate choice in red jungle fowl. Proc R Soc 259. Lond B 260:205–210. Kokko H, 1997. Evolutionarily stable strategies of age-dependent sex- Zuk M, Ligon D, Thornhill R, 1992. Effects of experimental manip- ual advertisement. Behav Ecol Sociobiol 41:99–107. ulation of male secondary sex characters on female mate prefer- Kokko H, Lindstro¨m J, 1997. Measuring the mating skew. Am Nat ence in red jungle fowl. Anim Behav 44:999–1006. 149:794–799. Zuk M, Thornhill R, Ligon JD, Johnson K, 1990a. Parasites and mate Kokko H, Lindstro¨m J, Alatalo RV, Rintama¨ki PT, 1998. Queuing for choice in red jungle fowl. Am Zool 30:235–244. territory position in the lekking black grouse (Tetrao tetrix). Behav Zuk M, Thornhill R, Ligon JD, Johnson K, Austad S, Ligon SH, Ecol 9:376–383. Thornhill NW, Costin C, 1990b. The role of male ornaments and Kokko H, Rintama¨ki PT, Alatalo RV, Ho¨glund J, Karvonen E, Lund- courtship behavior in female mate choice of red jungle fowl. Am berg A, 1999. Female choice selects for life time lekking perfor- Nat 136:459–473. mance in black grouse males. Proc R Soc Lond B 266:2109–2115. Lessels CM, Boag PT, 1987. Unrepeatable repeatabilities: a common mistake. Auk 104:116–121. Ligon JD, Kimball R, Merola-Zwartjes M, 1998. Mate choice by female red jungle fowl: the issues of multiple ornaments and fluctuating asymmetry. Anim Behav 55:41–50. Ligon JD, Thornhill R, Zuk M, Johnson K, 1990. Male-male compe- tition, ornamentation and the role of testosterone in sexual selec- tion in red jungle fowl. Anim Behav 40:367–373. Ligon JD, Zwartjes PW, 1995a. Female red junglefowl choose to mate with multiple males. Anim Behav 49:127–135. Ligon JD, Zwartjes PW, 1995b. Ornate plumage of male red jungle- fowl does not influence mate choice by females. Anim Behav 49: 117–125. Lindstro¨m J, 1994. Tetraonid population studies—the state of the art. Ann Zool Fenn 31:347–364. Moss R, Kolb HH, Marquiss M, Watson A, Treca B, Watt D, Glennie W, 1979. Aggressiveness and dominance in captive cock red grouse. Aggress Behav 5:59–84. Myhre G, 1980. Social status, external signals and colonic temperature in the captive willow grouse Lagopus lagopus lagopus. Ornis Scand 11:77–80. Nur N, Hasson O, 1984. Phenotypic plasticity and the handicap prin- ciple. J Theor Biol 110:275–297. Owens IPF, Short RV, 1995. Hormonal basis of sexual dimorphism in birds: implications for new theories of sexual selection. Trends Ecol Evol 10:44–47. Rintama¨ki PT, Alatalo RV, Ho¨glund J, Lundberg A, 1995a. Male ter-